COVID herd immunity: At hand or forever elusive?


With cases of COVID-19 either disappeared or rapidly diminishing from places like Wuhan, Italy, New York, and Sweden, many voices are speculating that herd immunity may have been reached in those areas and that it may be at hand in the remaining parts of the world that are still struggling with the pandemic.  Lockdowns should end—or may not have been needed to begin with, they conclude. Adding plausibility to their speculation is the discovery of biological evidence suggesting that prior exposure to other coronaviruses may confer some degree of immunity against SARS-CoV2, an immunity not apparent on the basis of antibody seroprevalence studies.

Opposing those viewpoints are those who dismiss the recent immunological claims and insist that rates of infections are far below those expected to confer immunity on a community. They believe that the main reason for the declining numbers are the behavioral changes that have occurred either under force of government edict or, in the case of Sweden, more voluntarily. What’s more, they remind us that the Spanish flu pandemic of 1918-1919 occurred in 3 distinct waves. In the summer of 1918 influenza seemed overcome until a second wave hit in the fall. Herd immunity could not possibly have accounted for the end of the first wave.

The alarmists may have a point.  However, recent history offers a more instructive example.

Until early 2015, epidemiologists considered Mongolia to be exemplary in how it kept measles under control. In the mid-1990s, the country instituted a robust vaccination program with low incidences of outbreaks, even by the standards of developed countries. In the early 2000s, it adopted a 2-step MMR immunization schedule and, after 2005, its vaccination rates were upwards of 95%. From 2011 through 2014, not a single case of the virus was recorded, leading the WHO to declare measles “eradicated” from Mongolia in November 2014.  

On March 18, 2015, however, a measles case occurred in the capital Ulaanbaatar, featuring a viral genotype similar to that of an outbreak in neighboring China. Within days, numerous other cases arose, marking the beginning of a multifocal epidemic across the country. A first wave of infections lasted until July and another one, begun in the fall of 2015, lasted until the spring of 2016. Taken together the outbreaks caused nearly 54,000 cases of measles and 140 deaths despite implementation of a program of supplemental immunizations instituted in the middle of 2015. Most of those affected were children less than 1 year of age who had not yet received any immunization, but 6% of the victims had received one or even 2 shots.  

What does this unfortunate event tell us about herd immunity?  A review paper titled “Herd Immunity: A Rough Guide,” written by Fine, Eames, and Heymann and published in 2011, gives important insights on the topic.

Herd immunity holds that if a sufficient number of people are immune to a contagious disease, then this immunity protects those who would otherwise be susceptible to it. The idea was first proposed in the 1930s when epidemiologists noted that the occurrence or disappearance of epidemics of common respiratory viruses, such as measles, mumps, or chickenpox, correlated with the presence or absence of a critical mass of immune individuals within a community. 

The widespread introduction of vaccines in the 20th century confirmed many examples of herd immunity, which occurs when an immune individual interferes with the chain of transmission of the contagious agent. Vaccination campaigns spurred the development of theoretical models to estimate “thresholds” of herd immunity based on estimates of transmissibility factor, or the famous R0. It is from such models a herd immunity threshold (HIT) estimates are derived and a HIT range of 50-80% has been proposed for COVID-19. In other words, models predict that 50-80% of the population must be infected and develop immunity to COVID-19 before the phenomenon of herd immunity can take effect.

Fine et al. describe in some detail the highly complex social and biological factors that bear on the modeling of herd immunity, and they warn against placing too much confidence in HIT estimates.  Despite growth in understanding and despite refinement in theoretical models, they believe that we ought to remain circumspect when considering the question of thresholds in practical terms: 

Managers [of vaccination programs] must be wary of target thresholds for vaccination, insofar as thresholds are based on assumptions that greatly simplify the complexity of actual populations. In most circumstances, the sensible public health practice is to aim for 100% coverage, with all the doses recommended, recognizing that 100% is never achievable, hoping to reach whatever is the ‘‘real’’ herd immunity threshold in the population concerned (emphasis mine).

The measles outbreak in Mongolia vindicates their cautionary advice. One might have felt justified into thinking that Mongolia possessed herd immunity on the eve of March 18, 2015 only to realize the next day that it didn’t. 

On reflection, the reason for the difficulty is simple. To be immune is a “negative” concept.  It cannot be positively established by empirical observation. It is only the opposite that can be demonstrated, i.e., that a person is infected. Of course, when knowledge about a particular infection is solid and biological markers of immunity are robust, the presumption of immunity can be made with more confidence, but it remains a presumption.  

Where does that leave us as far as COVID-19 lockdowns are concerned?  

Those who oppose lockdowns should be glad to see that Sweden and many other places may soon be “COVID-free.”  But invoking herd immunity as an explanation, persuasive as the biology may be, could be premature.  Arguments against lockdowns should look elsewhere for their grounding.

Conversely, opponents of lockdowns cannot be so sure that herd immunity has not occurred in some places.  Our understanding of the immunology and transmissibility of SARS-Cov2 is sufficiently sparse to allow very wide confidence intervals around any point estimate of herd immunity thresholds. The optimists could very well be right.

One last point.  At the end of their article, Fine et al. point out that most vaccination programs aim not to eradicate infectious disease but to reduce disease to a level that is “tolerable.” In a sense, then, our societal attitudes towards disease and death bear importantly on whether we perceive that herd immunity has been reached.

But in an age when health has been declared a common good and medicine has become irredeemably politicized, any case and any death can be deemed “intolerable” or weaponized for political aims.  That, more than anything else, may be what makes herd immunity particularly elusive.

Michel Accad is a cardiologist based in San Francisco and host of the podcast, The Accad & Koka ReportThis post originally appeared on his blog, Alert and Oriented here.

4 replies »

  1. Dr. Lundberg:

    From the paper cited in the article: “Mongolia maintained high measles vaccination coverage at > 95% since 2005” (Fourth paragraph done the top).

    I can only assume that means “95% of eligible children.”

    Thank you.

  2. “Those who oppose lockdowns should be glad to see that Sweden and many other places may soon be “COVID-free.”

    Only if you can tolerate all the dead people AND the overwhelming stress on the existing health care system, which impedes common access for other illnesses.

    “But in an age when health has been declared a common good…”

    Apparently not in the U.S.A.

  3. I absolutely love to read writing that just explores a topic, without an agenda, showing the viewpoint of both sides.

    It’s not ‘sitting on the fence’ or saying both sides are equally valid. It’s just being reasonable and discussing the facts as they are.

    Yes, there definitely is a time for persuasion and even fighting. But how refreshing when an intelligent, knowledgeable person can just inform and shed light, without trying to convince us of anything.

  4. ” In the early 2000s, it adopted a 2-step MMR immunization schedule and, after 2005, its vaccination rates were upwards of 95%. ”

    95% of what? All population? Children? What ages? Thanks for reply.